京都大学 / 新制・課程博士 / 博士(医科学) / 甲第23113号 / 医科博第124号 / 新制||医科||8(附属図書館) / 京都大学大学院医学研究科医科学専攻 / (主査)教授 萩原 正敏, 教授 篠原 隆司, 教授 上杉 志成 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Currently, human norovirus is the leading cause of acute gastroenteritis and accounts for most cases of foodborne illnesses in the United States each year. Due to its tissue culture inefficiency, studies of human norovirus have been crippled for more than forty years.Tulane virus (TV) stands out as a suitable surrogate of human norovirus given its high amino acid identity with human norovirus and its well-established cell culture system. It was first isolated from rhesus macaques (Macaca mulatta) in 2008 and identified as a novel Calicivirusrepresenting a new genus, Recovirus genus (Farkas et al., 2008). However, there are still unanswered questions about its infectious cycle and the essential factors for its infection. In this study, we have obtained a TV variant (the 9-6-17 strain) that has lost the binding ability to the B-type histo-blood group antigen (HBGA), which was proposed to be the receptor of both TV and human norovirus. In the first chapter, we outline how the sequence analysis,structural biology studies, and mutagenesis studies of the 9-6-17 TV strain have shed light on the interaction with its host cell receptor. To investigate the key residues for HBGA binding, we established the full-length infectious clone of the 9-6-17 TV strain. We present a highly selective transformation of serine 367, located in the predicted HBGA binding site, into a lysine residu e. Our results advance the understanding of genetic changes in TV required for adaptation to cell culture environments. Cryo-EM is an awarding winning technique that has been the greatest scientific breakthrough in recent years. It was awarded the Nobel Prize in Chemistry in 2017. Despite the technological advances of the direct electron detector and image processing software, several major roadblocks remain for high-resolution structure determination with cryo-EM. In the later chapters, we explored the most efficient way of using VPP to enhance image contrast, how to tackle the airwater interface problem by encapsulating target protein, how to reach a higher resolution by refining high order parameters, and the helical indexing problem in real space. These technical advances would benefit the whole cryo-EM community by providing convenient tools or insights for future directions.

Pore forming toxins (PFTs) are a large group of proteins found mainly in bacteria with some exceptions found in animals. They bind and form pores in their target membranes and form pores, which leads to cell death. Among these are cholesterol-dependent cytolysins (CDC), which require the presence of cholesterol to bind target membranes. Mitilysin (Mly), a protein of interest in this project, belongs to the CDC group of pore forming toxins. It is produced by the bacterium Streptococcus mitis, a pathogen closely related to Streptococcus pneumoniae, found in human oral cavity, which causes several diseases such as Viridans Group Streptococcal (VGS) toxic shock syndrome and endocarditis. Mly is a homologue of the toxin Pneumolysin, which is produced by S. pneumoniae. However, the mechanism of pore formation is not well known. The purpose of this project is to understand the mechanism of CDC pore formation, focusing on the key amino acid residues that are responsible for transitioning from Mly pre-pore to pore state. The findings will aid in the design of inhibitors of pore formation as potential anti-bacterial drug candidates. The major goal of the project was to determine the 3-dimensional (3D) structure of assembled Mly pore. Mly is expressed in E.coli and purified by Ni-NTA affinity chromatography. Pore formation is confirmed by a hemolysis assay and negative stain-transmission electron microscopy. Mly pores are vitrified, analyzed and imaged in a cryo-electron microscope. 2D images were processed to generate a 3D density map. However, our Mly pore 3D map was incomplete due to lack of 2D projection angles resulting from preferred orientation of pore particles during sample preparation. To overcome this problem, we aim to use DNA origami, which requires His-tagged Mly. We were able to determine that His-tagged Mly retains its pore formation ability.

Stem cells, with the ability to self-renew and differentiate into intended cell types, are a valuable tool for disease modeling and mechanistic study. CRISPR-Cas9 has been widely used for genome editing due to its high efficiency and convenience. However, CRISPR-Cas9 has large-deletion safety issues that dramatically restrict its applications. Wiskott-Aldrich syndrome (WAS) is an inborn immunological disorder caused by WASP deficiency. WASP functions in the nucleus, which may help to understand WAS pathology, are poorly defined. Pannexin 1 (PANX1) forms large plasma membrane pores to exchange intracellular small molecules with the extracellular environment and functions in inflammatory processes. The regulatory mechanisms of the PANX1 channel remain obscure. In this dissertation, I focused on mechanistic studies of CRISPR-Cas9 genome editing, and two immune disease relevant genes, WASP and PANX1 using stem cell-derived immune cells. We first found that CRISPR-induced large deletions (LDs) are predominantly mediated by the MMEJ repair pathway through statistical studies. Further, we found POLQ and RPA play vital roles in CRISPR-induced LDs. Modulation of POLQ and RPA can decrease CRISPR-induced LDs and increase HDR efficiency. Using three isogenic WAS iPSC models generated via gene editing, we successfully recapitulated WAS phenotypes, and for the first time, revealed that WASP regulates RNA splicing via epigenetically controlling the transcription of splicing factors and directly participating in the splicing machinery through a liquid-liquid phase separation process. We established a full-length human PANX1 (hPANX1) channel model via cryo-electron microscopy experiments and molecular dynamics simulation study, and found that hPANX1 channel is a homo-heptamer with both the N- and C-termini stretching deeply into the pore funnel. Functional studies of three selected residues support the new hPANX1 channel model and suggest the potential regulatory role of hPANX1 in pyroptosis upon immune responses. Overall, the mechanistic studies of WASP, PANX1 and CRISPR genome editing revealed new roles of WASP in regulating RNA splicing, new functional insights of PANX1 in pyroptosis, and uncovered two critical players POLQ and RPA in CRISPR-induced LDs.

No description available.

Strukturbestämning av proteiner är viktigt för att kunna förstå deras funktion och en snabbt utvecklande metod inom fältet är kryoelektronmikroskopi. Storleksbegränsningar förhindrar en bredare applikation av metoden eftersom små proteiner har för låg signal i förhållande till bakgrund för att kunna visualiseras som enstaka partiklar i elektronmiksoskopibilder. Hypotesen för projektet är att det är möjligt att avbilda väldigt små proteiner och kringgå den konventionella storleksbegränsningen genom att använda ett bärarprotein ((Putrescine Aminotransferase; YgjG) som kopplas till en affibody (Zwt) genom “helical fusion” och sedan binda ett litet målprotein till denna större struktur. Komplexet ska ge en tillräcklig storlek, symmetri och rigiditet för en lyckad elektronmikroskopi av bärare tillsammans med det lilla icke kovalent bundna målproteinet. För att karaktärisera den föreslagna bäraren, genomförs stabilitetstester genom CD, verifiering av inbindning av målproteinet i SPR, renhetsundersökning med SEC och slutligen kryoelektronmikroskopi för att testa konceptet. Det lilla målproteinet som kommer att avbildas i konceptstudien är en annan affibody (Z963), som i så fall skulle vara det minsta proteinet som någonsin har lösts med kryogenelektronmikroskopi. Resultaten visar att den undersökta tetramera-bäraren är väldigt stabil (Tm~ 85 oC) och kan tolerera en affibody-fusion med bibehållen bindning av multipla säten. Proteinet kan uttryckas rekombinant och renas till högt utbyte och bildar tetramerer även med fuserad affibody. De slutgiltiga resultaten från den kryoelektronmikroskopiska analysen inväntas fortfarande, men lovande griddar har skapats och en partikelselektion har gett klara 2-D klasser som också framhäver att det lilla målproteinet har bundit. Sammanfattningsvis har biofysikalisk karaktärisering indikerat att YgjG är en lovande bas för ett “imaging scaffold” och att preliminära enstaka-partikel kryoelektronmikroskopi analyser visar att den föreslagna strategin att undersöka små målproteiner är möjlig. / Determining structures of proteins is important to understand protein functions, and a rapidly evolving technique in this field is cryogen electron microscopy. However, size limitations are preventing wider applications of the technique because small proteins have poor signal to noise ratios and are not possible to distinguish in single-particle images. The hypothesis of this project is that it is possible to image very small proteins, bypassing the conventional size limitations of single-particle cryo-EM, by utilizing a carrier protein-scaffold (Putrescine Aminotransferase; YgjG) connected through helical fusion to an affibody (Zwt) that can bind to a small protein of interest. The complex provides a sufficient size, symmetry, and rigidity for successful electron microscopy also of the non-covalently bound small protein of interest. To characterise the proposed scaffold, thermal stability through CD, binding of target protein in SPR, purity through SEC and experiments towards proof-of-concept in cryo-EM will be performed. The small protein of interest to be imaged in the proof-of-concept setup is another affibody, called Z963, that would be the smallest protein ever solved with cryo-EM. The results show that the investigated tetrameric protein scaffold is a highly stable protein (Tm~85oC) that can tolerate affibody fusion with retained binding function of multiple sites. The protein can be recombinantly expressed and purified in high yield and forms tetramers also when fused to affibody. The cryo-EM results are still pending, but promising grids have been created and in an initial particle selection clear 2-D classes that also reveal the small bound protein of interest have been generated. To conclude, biophysical characterization indicates that YgjG is a promising base structure for an imaging scaffold and preliminary single-particle cryo-EM analyses show that the proposed strategy to investigate structures of small proteins of interest is feasible.

Small proteins are difficult to study at high resolution with single-particle cryo-electron microscopy (cryo-EM). In general, sample properties such as large size (> 80 kDa), symmetry and rigidity are key to utilize this technology. To facilitate structural studies of small proteins as well, using cryo-EM, this project aims to incorporate a photo-inducible cross-link in a large and symmetric scaffold that is amenable for study, and covalently bind small proteins of interest to this scaffold. The scaffold in this project consists of rabbit muscle aldolase (157 kDa in tetrameric state) with an engineered affibody affinity protein (7 kDa) attached to the N-terminus of each aldolase monomer via a rigid helix fusion. The affibody-domain of the scaffold will be cross-linked to small proteins of structural interest, with a focus on a model target consisting of a second affibody with affinity for the affibody displayed on the aldolase scaffold. Photoconjugation of the affibody Zwt was performed to crosslink both the Fc of IgG and the anti-idiotypic affibody Z963, revealing that a methionine acceptor in the target is preferable but not necessary for UV crosslinking using BPA. Binding of affibodies rigidly displayed on of the scaffold to targets such as affibodies and antibody fragments was demonstrated , using surface plasmon resonance (SPR). / Att studera små protein vid hög upplösning med enpartikelsrekonstruktion i kryo-elektronmikroskopi (kryo-EM) är utmanande. Generellt så krävs stora (> 80 kDa), symmetriska och stabila protein för att använda sig av kryo-EM. Med målet att möjliggöra strukturbestämning och strukturella studier av små protein, så ska detta projekt föra in en foto-aktiverad korslänk i ett stort och symmetriskt bärarprotein. Bäraren består av aldolas från kaninmuskel (157 kDa som tetramer) med en affibody (7 kDa) kopplad till N-terminalen av varje aldolas-monomer via en rigidt fuserad helix. Affibody-domänen av bärarproteinet kan bilda korslänkar till små protein vars struktur sedan kan studeras. Fokus i projektet är ett modellprotein som består av en annan affibody som binder den affibody i bäraren. Fotokonjugering av affibodyn Zwt utfördes för att skapa korslänkar till både Fc av IgG, samt den anti-idiotypiska affibodyn Z963, vilket påvisade att en metionin-mottagare i målproteinet är fördelaktigt för UV korslänkning med BPA, men inte ett krav. Affinitet av affibodies i bärarproteinet till målprotein såsom andra affibodies och antikroppsfragment påvisades.

Eine Hauptanwendung der Einzelpartikel-Analyse in der Kryo-Elektronenmikroskopie ist die Charakterisierung der dreidimensionalen Struktur makromolekularer Komplexe. Dazu werden zehntausende Bilder verwendet, die verrauschte zweidimensionale Projektionen des Partikels zeigen. Im ersten Schritt werden ein niedrig aufgelöstetes Anfangsmodell rekonstruiert sowie die unbekannten Bildorientierungen geschätzt. Dies ist ein schwieriges inverses Problem mit vielen Unbekannten, einschließlich einer unbekannten Orientierung für jedes Projektionsbild. Ein gutes Anfangsmodell ist entscheidend für den Erfolg des anschließenden Verfeinerungsschrittes. Meine Dissertation stellt zwei neue Algorithmen zur Rekonstruktion eines Anfangsmodells in der Kryo-Elektronenmikroskopie vor, welche auf einer groben Darstellung der Elektronendichte basieren. Die beiden wesentlichen Beiträge meiner Arbeit sind zum einen das Modell, welches die Elektronendichte darstellt, und zum anderen die neuen Rekonstruktionsalgorithmen. Der erste Hauptbeitrag liegt in der Verwendung Gaußscher Mischverteilungen zur Darstellung von Elektrondichten im Rekonstruktionsschritt. Ich verwende kugelförmige Mischungskomponenten mit unbekannten Positionen, Ausdehnungen und Gewichtungen. Diese Darstellung hat viele Vorteile im Vergleich zu einer gitterbasierten Elektronendichte, die andere Rekonstruktionsalgorithmen üblicherweise verwenden. Zum Beispiel benötigt sie wesentlich weniger Parameter, was zu schnelleren und robusteren Algorithmen führt. Der zweite Hauptbeitrag ist die Entwicklung von Markovketten-Monte-Carlo-Verfahren im Rahmen eines Bayes'schen Ansatzes zur Schätzung der Modellparameter. Der erste Algorithmus kann aus dem Gibbs-Sampling, welches Gaußsche Mischverteilungen an Punktwolken anpasst, abgeleitet werden. Dieser Algorithmus wird hier so erweitert, dass er auch mit Bildern, Projektionen sowie unbekannten Drehungen und Verschiebungen funktioniert. Der zweite Algorithmus wählt einen anderen Zugang. Das Vorwärtsmodell nimmt nun Gaußsche Fehler an. Sampling-Algorithmen wie Hamiltonian Monte Carlo (HMC) erlauben es, die Positionen der Mischungskomponenten und die Bildorientierungen zu schätzen. Meine Dissertation zeigt umfassende numerische Experimente mit simulierten und echten Daten, die die vorgestellten Algorithmen in der Praxis testen und mit anderen Rekonstruktionsverfahren vergleichen.

Electron crystallography of two-dimensional crystals is a structure-determination method well suited to the study of membrane protein structure-function. Two-dimensional crystals consist of ordered arrays of protein within reconstituted lipid bilayers, an arrangement that mimics the natural membrane environment. In this work we describe our recent progress in the use of this method with three different proteins, each providing a window into a separate paradigm in the electron crystallographic pipeline. Specific crystallization conditions for human leukotriene C₄ synthase (LTC₄S) have previously been determined, but our continued refinement of purification and crystallization has identified a number of additional parameters that greatly affect crystal size and quality, and we have developed a protocol to rapidly and reproducibly grow large, non-mosaic crystals of LTC₄S. The human gamma-glutamyl carboxylase (GGCX) has also been crystallized, but is sensitive to cryo-EM sample preparation conditions and we present here the successful reproduction of crystallization and refinement of cryo-EM sample preparation conditions. Lastly, we describe our crystallization screens with the Vibrio cholerae sodium-pumping NADH:ubiquinone reductase complex (Na⁺-NQR), and identify the factors critical to membrane reconstitution of the complex, a necessary first step towards crystallization. We also describe a semi-quantitative crystal screening protocol we have developed that provides quick and accurate method to assess two- dimensional crystallization trials, and discuss some general observations in optimization of membrane protein purification and two-dimensional crystallization for electron crystallography.

Les toxines de l’anthrax font partie de la famille des toxines A-B dans laquelle la moitié B se fixe à la membrane de la cellule permettant par la suite la translocation de la moitié A. Dans le cas de l’anthrax, la moitié B est représentée par le Protective Antigen (PA) et la moitié A par les deux protéines Edema Factor (EF) et Lethal Factor (LF). Après le recrutement par les récepteurs cellulaires (CMG2 et TEM8), PA s’organise en heptamère. Il peut fixer jusqu'à 3 ligands (EF et LF) avant d'être endocyté. Les modèles actuels de PA suggèrent que la baisse de pH à l’intérieur des endosomes permet un changement de conformation de la forme pré-pore vers la forme pore et que les ligands EF et LF passeraient au travers le pore pour entrer dans le cytoplasme. Cependant, le diamètre du pore est environ dix fois inférieur à celui des ligands (10 Å contre 100 Å). Un processus de folding/unfolding a été proposé mais demeure controversé. Afin d'identifier le processus de passage des facteurs EF et LF dans le cytoplasme, nous avons déterminé par cryo-microscopie électronique combinée avec l’analyse d’image les structures tridimensionnelles des complexes formés par PA et LF aux étapes prépore et pore. Par la suite, une étude complémentaire par dynamique moléculaire nous a permis de modéliser à haute résolution les différentes interactions qui ont lieu au sein du complexe. La structure 3D du complexe prépore combiné à 3 LF a été déterminée à une résolution de 14 Å. Nous avons aussi calculé une structure préliminaire du complexe pore également combiné à 3 LF Celles-ci n’ont jamais été résolues auparavant et leur connaissance permet d’envisager l’étude en profondeur du mécanisme infectieux de l’Anthrax in vivo. / The anthrax toxins are part of the A-B toxin family in which the B moiety binds to the cell membrane allowing subsequent translocation of the A moiety. In the case of anthrax, the B moiety consists of the Protective Antigen (PA), and the A moiety is composed of the two proteins Edema Factor (EF) and the Lethal Factor (LF). After being recruited by the cell receptors (CGM2 or TEM8), PA organizes itself into a heptamer. It can bind up to three ligands (either EF or LF) before being endocytosed. Current models suggest that the decrease of pH inside the endosomes allows a conformational change of PA from a prepore form to a pore form that allows the EF and LF ligands to pass through the pore and enter the cytoplasm. However, the pore diameter is about ten times smaller than the diameter of the ligands (10Å versus 100Å). A process of ligand folding / unfolding has been proposed, but remains controversial. To identify the mechanism by which EF and LF enter the cytoplasm, we have used cryo-electron microscopy and three-dimensional image analysis to determine the 3D structure of the PA-LF complexes in the pre-pore and pore conformations. Then, we used molecular dynamics to modelise at high resolution the different interactions that occur within the complex. The 3D structure of the pre-pore complex bound with three LF ligands has been determined at 14Å resolution. We also calculated a preliminary structure of the LF-bound pore complex. These structures have never been reported before. They provide the necessary information to study in depth the mechanism of anthrax infection in vivo.